Skip to content

Applied text mining techniques to compare the sentiment analysis about Donald Trump and Joe Biden during the last American elections after having extracted a sample of more than two million tweets.

Notifications You must be signed in to change notification settings

albertoparenti/Bsc-Thesis

Repository files navigation

Bachelor's Thesis

Introduction

The main topics covered in the following essay were analyzed on the occasion of the course "Statistical use of online databases" using tweets as examples concerning the main exponents of Italian political parties. This prompted me to apply the same methodology in a national reality where Twitter is used with higher frequency.

In the United States, the bluebird social network is a commonly used medium to express their opinions daily and to receive news. Considering data from Statista, a German database company specializing in market and data consumers, updated to June 2021, 187 million users access Twitter every day, 20% of them are based in the United States. To understand the importance of the information function of Twitter and its link with politics it is useful to highlight that the most represented professional category among those verified Twitter profiles is precisely that of journalists, 25% of the total (2015). It is curious to note that the profile more followed in the world is that of the President of the U.S.A. number 44, Barack Obama, with 130 million followers.

The analysis of the report arises from the desire to quantify public opinion American on the main candidates for the presidential elections on 4 November 2020. It was decided the study's key question: "How has the opinion of the American population changed towards the main candidates in the political elections between the election day (November 3, 2020)?" The analysis refers exclusively to Democratic Party candidate Joe Biden and Republican Party Donald Trump. To obtain an answer, sentiment analysis was applied to a dataset of tweets and R software was used.

The thesis is divided into 3 chapters: the first chapter deals with the applied methodology necessary to carry out the study. To be able to analyze the data clearly and exhaustive, it was necessary to apply cleaning and lemmatization operations to the sample downloaded from the Twitter social network. In this way, it was possible to realize the descriptive analysis discussed in the second chapter. After defining the type of tweet, yes both a quantitative and qualitative analysis is carried out through the analysis of hashtags and sentiment analysis. In the third chapter, we proceed to discuss the results obtained finally, from the analysis of the data and in the conclusions, a possible future study is hypothesized.

About

Applied text mining techniques to compare the sentiment analysis about Donald Trump and Joe Biden during the last American elections after having extracted a sample of more than two million tweets.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages